Otter.ai vs Google vs Rev: What Actually Happens to Your Recordings?

Otter.ai — cloud transcription with AI training by default

A cloud-based service. When you record a meeting or upload audio, it's uploaded to Otter's servers, transcribed on AWS infrastructure in the US, and stored there until you delete it.

What happens to your data

• Otter uses de-identified recordings and transcripts to train its own AI models by default — outside the Enterprise plan, there's no documented opt-out
• OtterPilot may take automatic screenshots during virtual meetings, which become part of the stored transcript
• Otter generates voiceprints (speaker-identification data) from recordings, listed as a sensitive data category in its Data Processing Addendum
• Connecting Google, Microsoft, or Zoom accounts gives Otter access to your calendar and contacts
• No end-to-end encryption is claimed in any primary document — Otter holds the encryption keys

The lawsuit

The consolidated class action alleges Otter Notetaker joins meetings as a "silent participant," records audio, takes screenshots, and stores everything without consent from all participants — then uses the contents to train AI models "for its own pecuniary gain." In August 2026, the judge denied Otter's motion to dismiss the CIPA, ECPA, and unjust enrichment claims, finding the plaintiffs had plausibly alleged Otter was "a third-party eavesdropper." These are allegations, not findings of fact — but the case proceeding past motion-to-dismiss means Otter will have to produce documents and answer questions under oath.

Otter says it doesn't sell personal information and that conversations belong to the user. Enterprise workspaces are opted out of training by default — Free, Pro, and Business accounts have no documented opt-out.

Google Workspace Voice Typing — free, but your audio leaves the building

Built into Google Docs. Click the microphone icon and your voice streams in real time to Google's cloud servers, gets processed by Google's speech-recognition AI, and the text appears in your document.

What happens to your data

• The complete audio of your dictation — including background conversations — is transmitted to Google in real time; there's no offline mode or on-device option
• Google's Workspace Terms of Service state voice recordings "may be retained" indefinitely for improving voice-recognition technology; Voice & Audio Activity is on by default for most accounts
• Voice data is used to train Google's machine-learning models — Workspace admins can't prevent this, it's built into the terms
• A Bloomberg investigation found human reviewers listen to recordings for quality assurance
• No published detail on whether audio or transcripts are encrypted in transit or at rest

Free and convenient, but for anyone handling privileged, medical, financial, or strategic information, audio leaving your device with indefinite retention is a real confidentiality risk.

Rev — better controls, but still cloud-based

A professional transcription service. You upload audio or video, and it's processed — by AI or by human transcriptionists — inside Rev's environment.

What's good

Rev states it never uses customer recordings to train third-party AI or LLMs, and doesn't share files with external vendors. Data is encrypted in transit and at rest. Administrators get audit logs of who opened which file and when. Backed by SOC 2 Type II, HIPAA-compliant workflows, CJIS support, and GDPR-compliant workflows.

What Rev doesn't say

Whether Rev trains its own (as opposed to third-party) AI models on customer data, specific retention periods, or whether automatic deletion exists outside enterprise contracts. If you request human transcription, Rev's transcribers will have access to your files as they work.

A strong choice for organizations needing compliance controls — but your audio still leaves your device and is processed on Rev's servers.